Multivariate statistical identification of human bladder carcinomas using ambient ionization imaging mass...

Area of Science:

• Analytical Chemistry
• Biomedical Engineering
• Oncology

Background:

• Accurate diagnosis of bladder cancer is crucial for effective treatment.
• Current diagnostic methods may require tissue processing, potentially altering molecular profiles.
• Mass spectrometry imaging offers a label-free approach to analyze tissue molecular composition.

Purpose of the Study:

• To develop and validate a diagnostic method for human bladder cancer using Desorption Electrospray Ionization Mass Spectrometry (DESI-MS).
• To correlate DESI-MS data with histopathological findings for accurate disease classification.
• To explore the potential of DESI-MS for clinical application in intact tissue diagnosis.

Main Methods:

• Analysis of twenty pairs of human bladder tissue samples (40 total) using DESI-MS.
• Characterization of glycerophospholipid (GP) mass spectral profiles for cancerous and normal tissues.
• Application of multivariate statistical analysis, including O-PLS-DA, for classification.
• Validation of the predictive model using independent training and validation sets.

Main Results:

• DESI-MS glycerophospholipid profiles effectively distinguished between cancerous and normal bladder tissues.
• The developed statistical model achieved high accuracy in classifying disease status.
• The model demonstrated a 5% error rate and 12% misclassification rate in validation.
• Generated synthetic images showed pixel-by-pixel disease classification that agreed with pathologist assessments.

Conclusions:

• DESI-MS combined with multivariate statistical analysis is a viable method for diagnosing human bladder cancer in intact tissue.
• The technique provides accurate, label-free molecular imaging that correlates with histology.
• This approach shows significant potential for clinical translation in cancer diagnostics.